EP2169820B1

EP2169820B1 - Alternating-current motor control apparatus

Info

Publication number: EP2169820B1
Application number: EP09169168A
Authority: EP
Inventors: Kozo Ide; Sadayuki Sato; Hideaki Iura; Shinya Morimoto
Original assignee: Yaskawa Electric Corp; Yaskawa Electric Manufacturing Co Ltd
Current assignee: Yaskawa Electric Corp
Priority date: 2008-09-26
Filing date: 2009-09-01
Publication date: 2010-12-29
Anticipated expiration: 2029-09-01
Also published as: ATE493792T1; JP5309838B2; US8143839B2; US20100079103A1; JP2010081726A; EP2169820A1; DE602009000495D1

Abstract

An alternating-current motor control apparatus includes a stator frequency computing unit configured to compute a stator frequency of a motor magnetic flux; a torque error computing unit configured to compute a torque error by using the motor magnetic flux, an estimated current, and a motor current; and a speed estimator configured to estimate a speed of the alternating-current motor by using the stator frequency and the torque error. The speed estimator includes a proportional controller configured to reduce the torque error to zero, and an adaptive filter configured to eliminate a high-frequency component of the torque error.

Description

BACKGROUND OF THE INVENTION

1. Field of the Invention

The present invention relates to an alternating-current (AC) motor control apparatus and an AC motor control method for performing torque control or speed control of an AC motor without using a position or speed sensor.

2. Description of Related Art

Methods for estimating the position and speed of an AC motor without using a position or speed sensor are roughly classified into methods in which the position and speed of an AC motor are estimated in accordance with a detected or estimated value of a motor induced voltage and methods in which the position and speed of an AC motor are estimated, by applying a high-frequency signal to the AC motor, in accordance with a detected value of a voltage or a current that depends on the inductance characteristic of the AC motor. The former method is suitable for driving an AC motor for which the inductance characteristic of the AC
motor is not available in advance. However, in the former method, in a case where the frequency at which the AC motor is driven is low, since the induced voltage is low, the signal-to-noise (S/N) ratio is reduced due to the influences of measured noise and the nonlinearity of characteristics of a driving circuit. Hence, a speed estimation error is increased.
For example, WO2002/091558 suggests a technique in which the speed of a motor is estimated, not directly in accordance with an induced voltage, but by estimating magnetic flux in accordance with a motor model, and at the same time, by estimating an error signal in accordance with an estimated value of magnetic flux and a deviation between a redundant estimated value of a current and a detected value of a current, using a proportional-plus-integral compensator that reduces the error signal to zero.
In addition, Japanese Unexamined Patent Application Publication No. 2003-319697 suggests a technique in which a gain computing unit is improved in such a manner that a gain of a deviation amplifier used for correcting the input of a motor model is properly output and the accuracy and responsiveness of speed estimation are thus improved while the reliability and responsiveness of speed estimation are taken into consideration.
Document JP08084500 proposes a speed sensorless vector controller for an induction motor using a rotor flux observer and an adaptive controller.
The invention is precisely defined in claim 1. The dependent claims recite advantageous embodiments of the invention.

BRIEF DESCRIPTION OF THE DRAWINGS

Exemplary embodiments of the present invention, given as non-limiting examples, will be described in detail based on the following figures, wherein:
Fig. 1 is a block diagram of an AC motor control apparatus according to a first embodiment;
Fig. 2 is a detailed block diagram of a speed estimator according to the first embodiment;
Fig. 3 includes illustrations for explaining filter characteristics of an adaptive filter according to the first embodiment;
Fig. 4A includes chart diagrams showing a case where the related art is applied;
Fig. 4B includes chart diagrams showing effects achieved in a case where an aspect of the present invention is applied;
Fig. 5 is a detailed block diagram of a speed estimator according to a second embodiment; and
Fig. 6 is a flowchart showing a control method performed in an AC motor control apparatus according to a third embodiment.

DESCRIPTION OF THE PREFERRED EMBODIMENTS

Hereinafter, embodiments of the present invention will be described with reference to the drawings.
Fig. 1 is a block diagram of an AC motor control apparatus I according to a first embodiment of the present invention.
The control apparatus I includes a current detector 102 for detecting three-phase currents (iu, iv, and iw) of a motor 101, and a three-phase/two-phase converter 103 for converting the three-phase currents (iu, iv, and iw) into detected two-phase currents (i_sα and i_sβ) in the rest system of coordinates.
The control apparatus I further includes a pulse width modulation (PWM) controller 104 for converting two-phase voltage commands (V*_sd and V*_sq) output from a vector controller 107 into three-phase voltage commands (V*u, V*v, and V*w) in the fixed system of coordinates by using a magnetic flux azimuth θ^ and applying the obtained three-phase voltage commands (V*u, V*v, and V*w) to the motor 101.
The control apparatus I further includes a phase computing unit 105 for computing the magnetic flux azimuth θ^ in accordance with an arctangent operation using estimated magnetic flux values (λ^_α and λ^_β) output from a motor model computing unit 109 and outputting the magnetic flux azimuth θ^ to the PWM controller 104 and a vector converter 106.
The control apparatus I further includes the vector converter 106 for performing coordinate conversion of the voltage commands (V*_sd and V*_sq) output from the vector controller 107 into two-phase voltage commands (V*_sα and V*_sβ) in the rest system of coordinates and outputting the two-phase voltage commands (V*_sα and V*_sβ) to the motor model computing unit 109.
The control apparatus I further includes the vector controller 107 for performing vector control of the motor 101 in the method described later; and a subtracter 108 for computing a difference (speed deviation Δωr) between a given speed command value ωr* and an estimated speed value ωr^ output from a speed estimator 114 and outputting the speed deviation Δωr to the vector controller 107.
The control apparatus I further includes the motor model computing unit 109 for computing estimated magnetic flux values (λ^_α and λ^_β) and estimated two-phase currents (i^_sα and i^_sβ) in accordance with the computation described later; and subtracters 110 and 111 for computing deviations (Δi_sα and Δi_sβ) between the estimated two-phase currents (i^_sα and i^_sβ) and the detected two-phase currents (i_sα and i_sβ) and outputting the deviations (Δi_sα and Δi_sβ) to a torque error computing unit 113.
The control apparatus I further includes a stator frequency computing unit 112, the torque error computing unit 113, and the speed estimator 114 for computing an estimated speed value ωr^ in the method described later, and drives the motor 101.
The vector controller 107 receives the speed deviation Δωr, a given magnetic flux command λr, and a magnetic flux component id and a torque component iq (not illustrated) of a motor current. The vector controller 107 performs speed control and current control in such a manner that the speed deviation Δωr is reduced to zero, and outputs the two-phase voltage commands (V*_sd and V*_sq) to the PWM controller 104 and the vector converter 106. Since methods for computing and controlling the magnetic flux component id and the toque component iq of a motor current and these commands are known, the explanation and illustration of these methods will be omitted.
The motor model computing unit 109 receives the two-phase voltage commands (V*_sα and V*_sβ) in the rest system of coordinates, and estimates magnetic flux values and currents in accordance with a mathematical model based on equations (1) and (2) as a motor model. The motor model computing unit 109 outputs the estimated magnetic flux values (λ^_α and λ^_β) to the phase computing unit 105, the stator frequency computing unit 112, and the torque error computing unit 113, and outputs the estimated currents (i^_sα and i^_sβ) to the subtracters 110 and 111 so that deviations (Δi_sα and Δi_sβ) can be calculated. In the following equations (1) and (2), vector notation is used, and voltage vector information represented as "V" in other parts of the description is represented as "u": $\frac{d}{dt} {\hat{i}}_{s} = a_{r 11} {\hat{i}}_{s} + (a_{r 12} - jρ {\hat{ω}}_{r}) \hat{λ} + b_{s} u_{s}$
$\frac{d}{dt} \hat{λ} = a_{r 21} {\hat{i}}_{s} + (a_{r 22} + j {\hat{ω}}_{r}) \hat{λ}$

where state variables are represented as a stator current vector: î _s = î _sα + j·î _sβ, a stator voltage vector: u_s = u _sα + j·u _sβ, and a magnetic flux vector: λ̂ = λ̂α + j·λ̂_β in the rest system of coordinates.
In addition, in the case of an induction motor, parameter definitions are as described below: $a_{r 11} = - \frac{1}{σ L_{s}} (R_{s} + R_{r}^{ʹ}), a_{r 12} = \frac{R_{r}^{ʹ}}{σ L_{s} M}, ρ = \frac{M}{σ L_{s} L_{r}}$
$a_{r 21} = \frac{M}{T_{r}}, a_{r 22} = - \frac{1}{T_{r}}, b_{s} = \frac{1}{σ L_{s}},$

where Rs represents a primary resistance, $R_{r}^{ʹ} = {(\frac{M}{L_{r}})}^{2} R_{r}$
represents a secondary resistance obtained by conversion on the primary side,
$Mʹ = \frac{M^{2}}{L_{r}}$
represents a mutual inductance obtained by primary conversion,
σLs represents a leakage inductance,
Ls represents a primary self-inductance,
Lr represents a secondary self-inductance,
$T_{r} = \frac{L_{r}}{R_{r}}$
represents a secondary time constant,
M represents a mutual inductance, and
ω̂ _r represents a rotator angular velocity.
Equations (1) and (2) are based on a continuous system. However, obviously, in the case of implementation, discretized equations may be used.
Next, the stator frequency computing unit 112, the torque error computing unit 113, and the speed estimator 114 will be sequentially described in detail.
The stator frequency computing unit 112 computes a stator frequency ω₀ in accordance with equation (3) by using the estimated magnetic flux values (λ^_α and λ^_β) estimated by the motor model computing unit 109: $ω_{0} = \frac{({\hat{λ}}_{α} \cdot \frac{d}{dt} {\hat{λ}}_{β} - {\hat{λ}}_{β} \cdot \frac{d}{dt} {\hat{λ}}_{α})}{{\hat{λ}}_{α}^{2} + {\hat{λ}}_{β}^{2}}$
A differential operation portion of equation (3) may be obtained by dividing a value obtained by subtracting the last magnetic flux value from the current magnetic flux value by a computation time, causing the computation result to pass through a low-pass filter, and eliminating a ripple portion generated in a sudden change.
The torque error computing unit 113 is provided to compute a difference between the estimated torque and the actual torque. However, since the actual torque cannot be directly measured, a torque error Δτ is computed by using the estimated magnetic flux values (λ^_α and λ^_β) estimated by the motor model computing unit 109 and the deviations (Δi_sα and Δi_sβ) computed by the subtracters 110 and 111, in accordance with equation (4): $Δ τ = {\hat{λ}}_{α} \cdot {Δ i}_{sβ} - {\hat{λ}}_{β} \cdot {Δ i}_{sα}$
Next, the speed estimator 114 will be explained. Fig. 2 is a detailed block diagram of the speed estimator 114. The speed estimator 114 includes a region discriminator 201, a cutoff frequency computing unit 202, a proportional controller 203, an adaptive filter 204, and an adder 205.
The region discriminator 201 is configured to perform conditional comparison of the stator frequency ω₀ and the torque error Δτ, and sets a coefficient g_i to 1 or 0. More specifically, in a case where the absolute value of the stator frequency ω₀ is smaller than or equal to a set value (about 1/200 of the rated driving frequency) and the absolute value of the torque error Δτ is equal to or greater than a set value (0.5 % of the rated torque), the coefficient g_i is set to 1. In a case where the above conditions are not met, the coefficient g_i is set to 0.
That is, in a case where the torque error Δτ increases in a region near the zero frequency, the coefficient g_i is set to 1.
The cutoff frequency computing unit 202 is configured to compute a cutoff frequency ω_i that is proportional to the torque error Δτ. By setting the conversion factor between the torque and the frequency to µ, the cutoff frequency ω_i is computed by using equation (5): $ω_{i} = μ \cdot |Δ τ|$
Note that the conversion factor µ should be set to within a range of about 1 to about 10 [rad/s] when the torque error Δτ is equal to the rated torque of the motor 101.
The adaptive filter 204 has the coefficient g_i and the cutoff frequency ω_i. The adaptive filter 204 receives the torque error Δτ, and computes a first estimated speed value ω^_r1 in accordance with equation (6): ${\hat{ω}}_{r 1} = \frac{Kiw \cdot (1 + g_{i} (ω_{i} - 1))}{s + g_{i} ω_{i}} Δ τ$
The filter characteristic of the adaptive filter 204 is set in such a manner that the adaptive filter 204 operates as a full integrator when the coefficient g_i is 0 and the filter bandwidth is increased to the cutoff frequency ω_i and the phase is changed from -90 degrees to 0 degree when the coefficient g_i is 1, as shown in Fig. 3.
The proportional controller 203 multiplies the received torque error Δτ by a gain Kpw in accordance with equation (7) and outputs the obtained value as a second estimated speed value ω^_r2: ${\hat{ω}}_{r 2} = Kpw \cdot Δ τ$
The adder 205 adds the first estimated speed value ω^_r1 to the second estimated speed value ω^_r2, and outputs the obtained value as the final estimated speed value ωr^.
As described above, the speed estimator 114 estimates the speed of the motor 101 by using the value obtained by adding the output of the proportional controller 203 configured to reduce the torque error Δτ to zero to the output of the adaptive filter 204 configured to eliminate a high-frequency component of the torque error Δτ.
Fig. 4A includes chart diagrams showing a case where the related art is applied. Fig. 4B includes chart diagrams showing effects achieved in a case where an aspect of the present invention is applied. Fig. 4A shows an estimated speed error and a phase error obtained, by using a known speed estimator employing proportional-plus-integral compensation, when the motor 101 is switched from normal rotation to reverse rotation in a rated load state of the motor 101. Fig. 4B shows an estimated speed error and a phase error obtained in a case where an aspect of the present invention is applied under the same conditions.
In the related art, near a region in which the speed is zero, an estimated speed error increases and the phase error accordingly increases. Meanwhile, according to an aspect of the present invention, both the estimated speed error and the phase error are reduced, and in particular, the speed error and the phase error are significantly reduced near the zero frequency, thus maintaining a reliable operation.
Since an AC motor control apparatus according to the first embodiment of the present invention is configured as described above, the operations and effects described below can be achieved.
Since the position and speed of a motor can be reliably estimated even in a region in which the driving frequency of the motor is low (including zero), torque control and speed control of the motor can be performed without using a position or speed sensor. Furthermore, the cutoff frequency of a filter used when a torque error is computed can be varied, and vibrations caused by the characteristics of the motor and a machine to which the motor is connected can be handled. Therefore, control instability can be reduced.
Fig. 5 is a detailed block diagram of a speed estimator 114' according to a second embodiment of the present invention. In the first embodiment, the speed estimator 114 in which the cutoff frequency computing unit 202 computes a cutoff frequency that is proportional to a torque error is used. Meanwhile, in the second embodiment, the speed estimator 114' is used instead of the speed estimator 114. That is, additional input signals are input to the speed estimator 114', and a cutoff frequency computing unit 502 of the speed estimator 114' computes a cutoff frequency that is proportional to a reactive power error Δq. As shown in Fig. 5, the speed estimator 114' further includes a region discriminator 501, a proportional controller 503, an adaptive filter 504, and an adder 505. Operations of the region discriminator 501, the proportional controller 503, the adaptive filter 504, and the adder 505 are the same as those of the region discriminator 201, the proportional controller 203, the adaptive filter 204, and the adder 205. Hence, the explanation of the region discriminator 501, the proportional controller 503, the adaptive filter 504, and the adder 505 will be omitted.
The cutoff frequency computing unit 502 will be explained. The cutoff frequency computing unit 502 computes a reactive power error Δq in accordance with equation (8) by using two-phase voltage commands (V*_sα and V*_sβ) in the rest system of coordinates and deviations (Δi_sα and Δi_sβ) computed by the subtracters 110 and 111, and computes a cutoff frequency ω_i in accordance with equation (9) by using a conversion factor µq between power and frequency: $Δ q = {V^{*}}_{sα} \cdot {Δ i}_{sβ} - {V^{*}}_{sβ} \cdot {Δ i}_{sα}$
$ω_{i} = μ_{q} \cdot |Δ q|$
Note that the conversion factor µq should be set to within a range of about 1 to about 10 [rad/s] when the reactive power error Δq is equal to the rated output of the motor 101.
Since, as with a torque error Δτ, the reactive power error Δq is caused by a speed estimation error, the adaptive filter 504 is capable of obtaining a first estimated speed value ω^_r1' in accordance with an operation similar to that of the adaptive filter 204 shown in Fig. 2.
As described above, the speed estimator 114' estimates the speed of the motor 101 by using the value obtained by adding an output of the proportional controller 503 that is configured to reduce the torque error Δτ to zero to an output of the adaptive filter 504 that eliminates a high-frequency component of the reactive power error Δq.
Since the speed estimator 114' is configured as described above in the second embodiment of the present invention, the operations and effects described below can be achieved.
Since the position and speed of a motor can be reliably estimated even in a region in which the driving frequency of the motor is low (including zero), torque control and speed control of the motor can be performed without using a position or speed sensor. Furthermore, the cutoff frequency of a filter used when a reactive power error is computed can be varied, and vibrations caused by the characteristics of the motor and a machine to which the motor is connected can be handled. Therefore, control instability can be reduced.
Fig. 6 is a flowchart showing a control method performed in an AC motor control apparatus according to a third embodiment of the present invention. A speed estimation method according to the third embodiment will be explained with reference to the flowchart of Fig. 6.
In step 1, motor magnetic flux values (λ^_α and λ^_β) and estimated currents (i^_sα and i^_sβ) are computed by using voltage commands (V*_sd and V*_sq) output from the vector controller 107 to the motor 101 and a magnetic flux azimuth θ^. This processing has been described above in the explanation of the motor model computing unit 109 in the first embodiment.
In step 2, a stator frequency ω₀ of the motor magnetic flux values (λ^_α and λ^_β) computed in step 1 is computed. This processing has been described above in the explanation of the stator frequency computing unit 112 in the first embodiment.
In step 3, a torque error Δτ is computed by using the motor magnetic flux values (λ^_α and λ^_β) and the estimated currents (i^_sα and i^_sβ) computed in step 1 and motor currents (i_sα and i_sβ) detected by using the current detector 102 and obtained by performing coordinate conversion. This processing has been described above in the explanation of the subtracters 110 and 111 and the torque error computing unit 113 in the first embodiment.
In step 4, a first estimated speed value ω^_r1 is computed by multiplying the torque error Δτ computed in step 3 by a proportional gain Kpw. This processing has been described above in the explanation of the proportional controller 203 in the first embodiment.
In step 5, a second estimated speed value ω^_r2 is computed by eliminating a high-frequency component of the torque error Δτ by using the stator frequency ω₀ computed in step 2 and a cutoff frequency ω_i determined in accordance with the torque error Δτ computed in step 3. This processing has been described above in the explanation of the adaptive filter 204 in the first embodiment.
In step 6, an estimated speed value ω^_r is computed by adding the first estimated speed value ω^_r1 computed in step 4 to the second estimated speed value ω^_r2 computed in step 5. The estimated speed value ω^_r is used for vector control and speed control performed in the vector controller 107 and the like in the first embodiment.
In the processing of step 5, as described in the second embodiment, a high-frequency component of the torque error Δτ may be eliminated by computing a reactive power error Δq by using voltage commands (V*_sd and V*_sq), estimated currents (i^_sα and i^_sβ), and motor currents (i_sα and i_sβ), and determining a cutoff frequency ω_i in accordance with the reactive power error Δq.
Since the control method performed in an AC motor control apparatus according to the third embodiment of the present invention is implemented, operations and effects similar to those of the first and second embodiments can be achieved.
According to the foregoing embodiments, the position and speed of a motor can be reliably estimated and torque control and speed control can be performed even in a region in which the driving frequency of the motor is low (including zero) by improving a speed estimator itself, without performing input correction of a motor model, unlike in the related art. Therefore, the present invention can be applied to general industrial machinery, in particular, to uses under circumstances in which a speed sensor cannot be used due to high temperature or high vibration.

Claims

An alternating-current motor control apparatus (I) including a pulse width modulation controller (104) for driving an alternating-current motor (101) by outputting a command voltage, comprising:
a motor model computing unit (109) configured to compute a motor magnetic flux and an estimated current of the alternating-current motor by using the command voltage;

a current detector (102) configured to detect a motor current flowing in the alternating-current motor;

a stator frequency computing unit (112) configured to compute a stator frequency of the motor magnetic flux;

a torque error computing unit (113) configured to compute a torque error by using the motor magnetic flux, the estimated current, and the motor current; and

a speed estimator (114, 114') configured to estimate a speed of the alternating-current motor by using the stator frequency and the torque error,

characterized in that the speed estimator estimates the speed of the alternating-current motor by using a value obtained by adding an output of a proportional controller (203, 503) configured to reduce the torque error to zero to an output of an adaptive filter (204, 504) configured to eliminate a high-frequency component of the torque error, the adaptive filter (204, 504) having a coefficient (g_i) determined in accordance with a cutoff frequency (ω_i) associated with the torque error (Δτ), the torque error (Δτ), and the stator frequency (ω_o).
The alternating-current motor control apparatus according to Claim 1, wherein the cutoff frequency is proportional to the torque error.
The alternating-current motor control apparatus according to Claim 1, wherein a cutoff frequency is computed in accordance with a reactive power error computed by using the command voltage, the estimated current, and the motor current, and the adaptive filter has a coefficient determined in accordance with the cutoff frequency, the torque error, and the stator frequency.
The alternating-current motor control apparatus according to Claim 3, wherein the cutoff frequency is proportional to the reactive power error.
The alternating-current motor control apparatus according to any one of Claims 1 to 4, wherein the adaptive filter operates as an integrator when the coefficient is 0, and operates as a primary delay filter when the coefficient is 1.